[ibis-macro] Re: AMI_Flows_6.pdf for today's ATM teleconference

  • From: "Muranyi, Arpad" <Arpad_Muranyi@xxxxxxxxxx>
  • To: "IBIS-ATM" <ibis-macro@xxxxxxxxxxxxx>
  • Date: Wed, 21 Apr 2010 11:39:01 -0700

Ambrish,
 
A couple of comments:
 
"But I strongly disagree that inside one model Init and GetWave can
provide different approximations of the same algorithm"
 
It seems that the usage of the word "can" is not proper in the middle.
I feel "should" would have been more appropriate in the context.  As
far as I understand, according to the current spec this "CAN" be done
aside from the error in reversing steps 4 and 5 pointed out on pg 1 in:
 
http://www.vhdl.org/pub/ibis/macromodel_wip/archive/20091009/arpadmurany
i/AMI%20flows:2009%20Sept%2029%20proposal%20-%20fixed/AMI_Flows_2fixed.p
df
 
Also, this does not imply "introducing double-counting" if done
properly, which was the whole point of the "final flow" we got
done in November last year, using an "... option to switch the
model between the statistical and non-linear mode".
 
I am sure there are many ways to skin the cat (the flow), but the
point I am trying to make is that we spent about three months last
fall to develop a flow that seemed acceptable to most of us, and
now we are practically starting the discussions over from scratch
because we are rehashing the same old arguments on what the
original intent was or wasn't, whether we should have LTI and
non-LTI algorithms in the same model, etc...
 
The November flow solved a bunch of problems, why do we want to give
ourselves more work now to undo all that and start from scratch just
because by solving the problems we opened the door to a few new
useful things?  It seems to be insane to me to undo all the work we
have put into the flow knowing that they will most likely end up in
a separate BIRD anyway, when we already have it ready to go now.
 
Arpad
======================================================================
 


________________________________

From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Ambrish Varma
Sent: Wednesday, April 21, 2010 12:30 AM
To: IBIS-ATM
Subject: [ibis-macro] Re: AMI_Flows_6.pdf for today's ATM teleconference


"The comments we are getting from Kumar and Ken indicates that they
would prefer to further simplify the flow, not allowing for the dual
nature of LTI and non-LTI coexisting together in the same model."
 
But we already have that intention in sections 2.1 and 2.2. and other
parts of the spec - ie - no getwave if model is linear and can be
described in init.
 
I stumbled upon an email that Danil wrote to Fangyi last year that says
it pretty nicely.
 
---
Fangyi, 
 
Could you please clarify it? I agree that it might be reasonable to put
the linear part of the model in Init and non-linear part in Getwave, so
that together they characterize the model (statistical simulator uses
only the linear part, while the pattern-dependent always uses both).
 
But I strongly disagree that inside one model Init and GetWave can
provide different approximations of the same algorithm (i.e. introducing
double-counting), where statistical simulator uses Init and
pattern-dependent simulator uses GetWave. I believe this behavior should
be prohibited, since it makes the flow more complicated, and we can
easily achieve the same result providing two different models (or having
internal option to switch the model between the statistical and
non-linear mode). 
 
If we have this simple rule (non-linear simulator we always uses Init
and Getwave), the behavior of the EDA does not depend on the fact
whether GetWave exists or not, and GetWaveExists flag becomes
unnecessary (if the Simulator at some point figures out there is no
GetWave, it just does not use it).
 
Are we on the same page here?
 
Best,
 
Danil
 
---
 
Another important data point is the default value of Use_Init_Output. It
was set to be true because we did not expected double counting by having
same or similar algorithms in Init and Getwave.
 
Thanks,
Ambrish.
 
 
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Wednesday, April 21, 2010 12:37 AM
To: IBIS-ATM
Subject: [ibis-macro] Re: AMI_Flows_6.pdf for today's ATM teleconference
 
That's exactly what I presented in the flow diagram I
prepared for today, but we are still not in agreement.
 
The comments we are getting from Walter indicate that
the entire flow is flawed and he suggests to go back
to the flow we finished in November.  The comments we
are getting from Kumar and Ken indicates that they
would prefer to further simplify the flow, not allowing
for the dual nature of LTI and non-LTI coexisting
together in the same model.
 
How shall we resolve this disagreement?
 
Arpad
========================================================
 
 
 
 
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Ambrish Varma
Sent: Tuesday, April 20, 2010 11:27 PM
To: IBIS-ATM
Subject: [ibis-macro] Re: AMI_Flows_6.pdf for today's ATM teleconference
 
Arpad,
What Kumar, Ken and I are trying to say (along with a few others, I
believe) is that the flow, as described in section 2.1, 2.2 and 2.3 of
chapter 10, works, and the only clarification (correction) that is
needed is the Tx_Getwave issue.
Adding any other flow is unnecessary and will add to more confusion.
 
It will also go against the main goal of the ATM committee as set by
you.
 
Regards,
Ambrish.
 
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Tuesday, April 20, 2010 11:23 PM
To: IBIS-ATM
Subject: [ibis-macro] Re: AMI_Flows_6.pdf for today's ATM teleconference
 
Kumar,
 
I heard this from you and Ken and others before, and I
have already responded to it more than once that making
a change of this nature is a "new feature" as far as the
specification is concerned, so I am simply not going
to entertain the idea, at least not with this BIRD and
this discussion.
 
Arpad
=========================================================
 
 
 
-----Original Message-----
From: ckumar [mailto:ckumar@xxxxxxxxxxx] 
Sent: Tuesday, April 20, 2010 8:59 PM
To: Muranyi, Arpad
Cc: IBIS-ATM
Subject: Re: [ibis-macro] Re: AMI_Flows_6.pdf for today's ATM
teleconference
 
Sigrity is of the view models employing getwave preferabley should be
using init only for parameter passing. This will greatly simplify the
flow,
minimize confusion  and more importantly will not potentially result in
the
problem of the same model yielding two different results.
 
A useful analogy is getwve vs init similar as silicon vs behavior
circuit
models
---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@xxxxxxxxxxxxx
  Subject: unsubscribe
 
---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@xxxxxxxxxxxxx
  Subject: unsubscribe
 
---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@xxxxxxxxxxxxx
  Subject: unsubscribe

Other related posts:

  1. Home
  2. ibis-macro
  3. Archive
  4. 04-2010